Re: 2nd attemt at reviving the filesystem limit discussion.

[Thomas Bushnell, BSG]

nisse@lysator.liu.se (Niels Möller) writes:

> > This breaks network transparency, and is a bug in the MK8x versions of
> > the kernel, that had to be fixed in the NORMA versions.
> 
> There must be some saner way to solve that problem. For integer
> arguments, the network stubs could integers in some bignumish way, and
> the receiving stub would convert it to the local machine word or
> return an int-is-too-big-error message on overflow. (To get it right
> one must treat signed and unsigned integers differently, of course).

Those are certainly good ways to fix the problem.

> If it's too late to fix Mach, I hope we can at least do better for L4.

Indeed!  They just made a little mistake in Mach's interfaces,
entirely forgivable.  They hadn't yet realized the true facts of what
network transparency requires, and so they missed some of these
things, especially since they didn't actually send those messages over
the network for a very long time.

The point is that there should be no machine-dependencies in the
actual message formats, of course.